- Email: [email protected]
The role of carbohydrate in human choriogonadotropin (hCG) action. Effects of N-linked carbohydrate chains from hCG and other glycoproteins on hormonal activity
263
Molecular and Cellular Enakcrinology, 70 (1990) 263-272 Elsevier Scientific Publishers Ireland, Ltd.
MOLCEL 02279
The role of carb.ohydrate in human choriogonadotropin
(hCG) action
Effects of N-linked carbohydrate chains from hCG and other glycoproteins on hormonal activity N. Rao Thotakura ‘, Bruce D. Weintraub
’
and Om P. Bahl 2
’ Molecular, Cellular and Nutritional Endocrinology Branch, NIDDK NIH, Bethesda, MD, U.S.A., and * Department of Biological Sciences, State Universiry of New York at Buffalo, Amherst, NY, U.S.A. (Received 11 January 1990; accepted 7 February 1990)
Key words: Glycoprotein; Gonadotropin;
Carbohydrate chain; Receptor binding; Hormone; N-linked oligosaccharide;_(Rat
ovary)
Deglycosylation of gonadotropins and thyrotropin results in a major loss of hormonal bioactivity, while not impairing receptor-binding activity. However, a direct role of the glycan moieties in hormonal signal transduction has not been demonstrated. The addition of carbohydrate chains together with the deglycosylated hormone does not restore the hormonal activity. In contrast, glycopeptides were found to inhibit human choriogonadotropin (hCG)-stimulated adenylyl cyclase activity and hCG binding to its receptor. An inhibition of hCG-stimulated adenylyl cyclase activity but not hCG binding to receptor by glycopeptides specifically from hCG, has previously been reported as a lectin-like membrane component has been implicated in hCG action. In the present study we have shown that glycopeptides and oligosaccharides prepared from hCG, transferrin, fetuin, q-acid glycoprotein and ovalbumin inhibit the binding of hCG to its receptor. The inhibition was also observed with a highly purified preparation of the receptor, thus suggesting a lack of involvement of other lectin-like membrane components as previously proposed. We suggest that a lectin-like interaction with the hormone, if any, involves the receptor itself. Adenylyl cyclase activity stimulated by hCG, isoproterenol or forskolin was inhibited by oligosaccharides, indicating a non-specific interaction. Our results suggest that Asn-linked oligosaccharide chains from various glycoproteins perturb hCG-receptor interactions through a putative carbohydrate binding site on the receptor.
Introduction
The glycoprotein hormones, choriogonadotropin (CG), lutropin, follitropin and thyrotropin are
Address for correspondence: N. Rae Thotakura, Building 10, Room 8D14, NIH, Bethesda, MD 20892, U.S.A. Supported in part by USPHS grant HD 08766.
composed of two dissimilar, non-covalently linked subunits, cu and 8. The common a-subunit of glycoprotein hormones contains two Asn-linked carbohydrate chains, while the /&subunit contains one or two Asn-linked units depending upon the hormone (Pierce and Parsons, 1981). The majority of these carbohydrate chains are biantermary complex type with minor populations of mono- and triantennary chains (Kessler et al., 1979; Kobata, 1984).
264
The importance of carbohydrates in the function of glycoprotein hormones has been well documented (Manjunath and Sairam, 1982; Bahl et al., 1984; Ryan et al., 1987). Enzymatic (Moyle et al., 1975; Goverman et al., 1982) or chemical deglycosylation of hCG and other glycoprotein hormones results in a loss of their post-receptor biological activity (Thotakura and Bahl, 1982; Bahl and Kalyan, 1983; Kalyan and Bahl, 1983). Data from our laboratory have shown that incubation of ovarian membrane hCG receptor with deglycosylated hCG and glycopeptides from hCG, cri-acid glycoprotein or ovalbumin did not restore the adenylyl cyclase-stimulating activity of the hormone (unpublished). The exact mechanism of how the carbohydrate chains on hCG bring about the activation of the hormone-occupied receptor is not known. It is currently believed (Kobata, 1984; Rademacher et al., 1988) that the carbohydrate chains of hCG bind to a lectin-like membrane component and thereby mediate the activation of adenylyl cyclase. These conclusions stemmed from a report (Calvo and Ryan, 1985) showing an inhibition of hCG-stimulated adenylyl cyclase but not receptor binding of HCG by glycopeptides specifically from hCG. If this effect is due to the oligosaccharides, it is hard to conceive that hCG glycopeptides show specific effects, since several other glycoproteins contain structurahy similar oligosaccharide chains identical to those on hCG. Moreover, in our studies, we have found inhibition of adenylyl cyclase as well as receptor binding of hCG by glycopeptides from several other glycoproteins. Since the role of carbohydrate in glycoprotein hormone action is an important question, we have reevaluated the effects of Asn-linked carbohydrate chains on hCG action using glycopeptides as well as oligosaccharides from hCG and various other functionally unrelated glycoproteins like transferrin, fetuin, cu,-acid glycoprotein and ovalbumin containing different types (bi-, triand tetra-antemrary as well as high marmose type) of carbohydrate chains. Materials and methods Oligosaccharides
and glycopeptides
Oligosaccharides from hCG, transferrin, fetuin, ovalbumin and cy,-acid glycoprotein were pre-
pared by hydrazinolysis according to the procedure of Takasaki et al. (1982). Lyophilized samples of the glycoproteins were hydrolyzed with anhydrous hydrazine for lo-12 h at 110°C in a sealed tube. After drying and removing the hydrazine, the samples were dissolved in saturated NaHCO, and N-acetylated by the addition of acetic anhydride. The oligosaccharides were purified on a Dowex 50 X 8-H+ column and then on a Bio Gel P-4 column. An aliquot of each of the oligosaccharides was reduced with NaB3H, in 0.05 N NaOH for 4 h at 30’ C. The samples were then neutralized with 1 N CH,COOH, purified by Dowex chromatography and the borohydride removed by repeated evaporation with methanol. After N-acetylation, the oligosaccharides were purified on a Bio Gel P-4 column (Yamashita et al., 1982). To assess their purity, an aliquot of each of the oligosaccharides was reduced with NaB3H, (Takasaki et al., 1982), chromatographed on Silica Gel 60 (Merck) in n-propanol/acetic acid/water (3 : 3 : 2) and autoradiographed. All the oligosaccharides prepared showed a single spot on thin-layer chromatography in the above system. The structures of all these oligosaccharides have previously been established (Kobata, 1984). Glycopeptides from hCG, ovalbumin and (piacid glycoprotein were prepared by extensive pronase digestion as described (Kessler et al., 1979). Briefly, the glycoproteins were reduced with dithiothreitol and the disulfide groups were carboxamidomethylated with iodoacetamide. The glycoprotein samples were digested with pronase for 96 h at 37’ C in an atmosphere of nitrogen with 1% by weight of the enzyme added at 24 h intervals. The resulting peptides were separated on Sephadex G-SO(fine) and the fraction were monitored for carbohydrate using phenol-sulfuric acid reaction. The glycopeptide fractions were pooled, desalted and lyophilized. Receptor binding assays
Receptor binding assays for hCG were performed using superovulated rat ovary membranes (Thotakura and Bahl, 1982). Ovaries were homogenized in 50 mM Tris-HCI, pH 7.2 containing 1 mg/ml bovine serum albumin and centrifuged at 10,000 X g for 20 min. The membrane pellet was resuspended in the same buffer and incubated
265
with 100,000 cpm 1251-hCG (35-50 $i/pg, iodinated by the chloramine T method) for 1 h at 37 o C in the presence or absence of various concentrations of unlabeled hCG and/or glycopeptides and oligosaccharides. Bound label was counted after separating the membranes by centrifugation for 15 min at 1500 X g. In experiments with pure receptor, an aliquot of affinitypurified, homogenous soluble receptor preparation was incubated in 50 mM Tris-HCl, pH 7.4 containing 20% glycerol, with labeled hCG for 18 h at 4 o C, and the bound label was separated by precipitation with poly(ethylene glycol) 8000, at a final concentration of 10% (Bruch et al., 1986). Beta-adrenergic receptor binding was assayed by incubat~g membranes prepared as above with ‘251-labeled iodocyanopindolol (ICYP) and the effectors in an incubation buffer containing 25 mM Tris-HCl, pH 7.4,l mM EDTA, 1 mg/ml bovine serum albumin, 0.1 mM ascorbic acid, 5 mM MgCl, for 30 min at 32” C. The membranes were collected by quickly filtering on GF/C filters under vacuum, washed twice with buffer and counted for bound radioactivity. In experiments testing the direct binding of carbohydrate chains to the membranes, two ovaries where homogenized in 10 ml of 50 mM Tris-HCl, pH 7.2 containing 1 mg/ml bovine serum albumin. Oligosaccharides labeled as described above, in 0.2 ml were incubated with 0.5 ml ovary membranes in 50 mM Tris-HCl, pH 7.2 containing 1 mg/ml bovine serum albumin for 1 h at 37” C. After separation of the membranes by centrifugation, the pellet was solubilized and counted in a liquid scintillation counter. Adenylyl cyclase assay Ovary membranes as prepared
above were incubated with hCG and effecters in an assay mixture containing 50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mg/ml bovine serum albumin, 2 mM ATP, 5 mM MgCl,, 10 PM GPP(NH)P, 0.2 mM SQ20009, 10 mM creatine phosphate, 50 U/ml tireanne kinase and 0.5 &i of [o-32P]ATP in a total volume of 0.1 ml. Incubations were carried out for 15 min at 34OC and terminated by the addition of excess unlabeled ATP, about 10,000 cpm of ‘H-cyclic AMP and 0.1% sodium dodecyl sulfate. The cyclic AMP formed was determined
by sequential chromatography on Dowex alumina columns (SaIomon, 1979).
and
Resuits Effect of glycopeptides and oligosaccharides on the binding of hCG to its receptor
We have previously observed that incubation of ovary membranes with deglycosylated hCG and glycopeptides from hCG, ovalbumin or or-acid glycoprotein failed to restore the adenylyl cyclase-stimulating activity of the hormone. However, the glycopeptides were found to inhibit adenylyl cyclase activity stimulated by hCG and further, in receptor binding studies, the glycopeptides were found to inhibit hCG binding to its receptor. As shown in Fig. 1, both glycopeptides and oligosaccharides, not only from hCG, but from a,-acid glycoprotein, ovalbumin and transferrin inhibited ‘*%hCG binding to its receptor on superovulated rat ovary membranes, This effect was observed at very high ~n~ntrations of the glycopeptides and oligosaccharides, three to four orders of magnitude higher than that of hCG required to cause similar displacement. The difference in the potencies of the various glycopeptides and o~gosa~h~des is probably related to their different carbohydrate contents. Preincubation of the ovary membranes with the glycopeptides also showed identical inhibition curves (not shown). Since the inhibition was noted with both glycopeptides as well as oligosaccharides, and peptide fractions with no ~bohy~ate chains are without any effect (not shown), we conclude that the observed effects are due to the carbohydrate moieties and not the peptide component of the glycopeptides. Since the inhibition of hCG binding was caused by the ~bohydrate, further experiments were performed to determine if monosaccharides could cause the same effects. Galactose, marmose and N-acetyl glucosamine at concentrations up to 1 mg and sialic acids up to 250 pg per assay did not have any effect on the binding of hCG to its receptor in ovary membrane receptor binding assay (Fig. 2). These data indicate that for the inhibitory effect of carbohydrates on hCG binding, a simple monosaccharide structure is not sufficient and rather, some common structural fea-
266
-
Further studies were performed to validate the involvement of any lectin-like membrane component in the inhibition of hCG to its receptor by oligosaccharides, as previously proposed (Calvo and Ryan, 1985). We have previously purified hCG/lutropin receptor from rat ovary (Bruch et al., 1986), and using this preparation, binding studies were performed in the presence of glycopeptides and oligosaccharides. Fig. 4 shows that the glycopeptides and oligosaccharides inhibited hCG binding to pure receptor as they had in membrane receptor assays. This shows that the effect of glycopeptides and oligosaccharides on hCG binding to its receptor does not require the integrity of membrane and does not involve any other membrane component.
L,.
0 0
“>..
..
““.“n
..I
10 hCG,
Glycopeptide
or
1
120
-i
100
1000
ng/tube
ollgosaccharlda,
ugitube
Fig. 1. Effect of glycopeptides and oligosaccharides on hCG binding to membrane receptor. Bat ovary membranes were incubated with ‘251-hCG ( -100,000 cpm) and the indicated concentrations of unlabeled hCG or various glycopeptides and oligosaccharides in 50 mM Tris-HCl containing 1 mg/ml bovine serum albumin in a total volume of 1 ml for 1 h at 37OC. Membranes were separated by centrifugation and the bound label quantitated. Binding in the absence of any effector was usually about 20% of the total counts added and is taken as 100%. The data shown are from a typical experiment re peated at least 3 times. (+) hCG; (A) cx,-acid glycoprotem glycopeptide; (W) ovalbumin glycopeptide; (0) hCG glycopeptide; (0) hCG oligosaccharide; (A) transferrin oligosaccharide; (0) ovalbumin oligosaccharide.
100
q q 0
Control GA Man GlcNAc Slalic Acids $“:3”$gp
GP
I
s E
80
z *0 2 i
60
z I : 40
ture(s) present in high mannose and complex type oligosaccharides is required. Another interesting observation was the ability of the glycopeptides to displace the hormone already bound to the receptor. The binding of hCG to its receptor is of high affinity and is reversible very slowly over a period of several hours as has been demonstrated previously (Bruch et al., 1986). However, when hCG was allowed to achieve equilibrium binding by incubation with the receptor for 1 h at 37” C, and further incubated with glycopeptides at concentrations comparable to those used in Fig. 1, a very rapid reversal of the hormone binding was observed (Fig. 3). Again, monosaccharides were without effect (not shown).
20
0
Fig. 2. Effect of monosaccharides on hCG binding to receptor. Bat ovary membranes were incubated with “‘1-hCG and the indicated monosaccharides under conditions as described in Fig. 1. Gal, Man and GlcNAc were each present at 1 mg/tube and sialic acids at 250 pg. Inhibition in the presence of hCG and glycopeptides is also shown for comparison. cut-Acid glycoprotein glycopeptide (cd Acid gp GP) and ovalbumin glycopeptide (Oval GP) were added at 500 pg each per tube. Values are mean f SE of triplicate determinations.
261
w GL.P ILP
25
and its receptor or to a non-specific effect on membrane proteins, we took advantage of the presence of /3-adrenergic receptors in the oy2y (Thotakura and Bahl, 1982). Binding of Icyanopindolol to ovary membranes was measured in the presence of different concentrations of ovalbumin oligosaccharides and q-acid glycoprotein glycopeptides. As shown in Fig. 5, the oligosaccharides and glycopeptides did not affect j?adrenergic receptor binding in the same membrane preparations in which hCG binding was inhibited. Moreover, hCG and bovine thyrotropin at 1 pg/O.S ml concentration were without any effect. These results indicate that the effect of the carbohydrate units is specific to hCG binding to its receptor.
0
Fig. 3. Effect of glycopeptides on the dissociation of receptorbound hCG. Ovary membranes were incubated with 12?-hCG for 1 h during the first incubation and the incubation was continued for an additional 1 h after adding 500 cg each of q-acid glycoprotein glycopeptide (al Acid gp GLP) and ovalbumin glycopeptide (Gvalb GLP), 1 pg hCG or none in the control. The amount of label bound in control is taken as 100%. Values are mean f SE of triplicate determinations.
One possible explanation for the inhibition of hCG-receptor binding by oligosaccharides is that the receptor has a low affinity binding site for the carbohydrate, which is separate from the polypeptide binding site and the binding of glycopeptides to this low affinity site inhibits the binding of hCG-carbohydrate, and thereby displaces hCG from its receptor. To test this hypothesis, direct binding of 3H-labeled oligosaccharides from hCG and ovalbumin to ovarian receptor was studied. We were unable to demonstrate any significant specific binding of the labeled oligosaccharides to ovary membranes as shown in Table 1. Similar results were obtained when glycopeptides labeled with 12? using Bolton-Hunter reagent were used (not shown). Effect of glycopeptides and oligosaccharides on j3adrenergic receptor binding To determine whether the effect of glycopeptides and oligosaccharides was specific to hCG
OF““‘i
.“‘lrl hCG,
Glycopeptide
or
“’
’
100
l(
IO
ngltube
ollgosaccharlde,
vgmba
Fig. 4. Effect of glycopeptides and oligosaccharides on hCG binding to purified soluble receptor. Affinity-purified soluble ovarian hCG receptor was incubated with the indicated concentrations of unlabeled hCG or glycopeptides or oligosaccharides for 18 h at 4’ C. Receptor-bound label was separated by precipitating with poly(ethylene glycol) as described under Materials and Methods. Values are plotted as percent of label bound in the absence of any effector. Data shown are from a representative experiment repeated at least 3 times. (+) hCG; (0) q-acid glycoprotein oligosaccharide; (0) fetuin oligosaccharide; (A) q-acid glycoprotein glycopeptide; (m) ovalbumin glycopcptide.
268
Effect of oligosaccharides ity
on adenylyl cyclase activ-
To determine whether the inhibitory effects of the glycopeptides and oligosaccharides on receptor binding can be observed at the post-receptor actions of the hormone, we have assayed adenylyl cyclase activity of ovary membranes stimulated by hCG, in the presence of various preparations of oligosaccharides. Fig. 6A shows a dose-dependent inhibition of adenylyl cyclase activity stimulated by 100 ng hCG by oligosaccharide prepared from hCG. Further studies showed (Fig. 6B) that this effect was not specific to hCG-oligosaccharides. Oligosaccharides from fetuin, a,-acid glycoprotein and ovalbumin were also able to inhibit hCGstimulated adenylyl cyclase activity. Moreover, the activity stimulated by 10 PM forskolin and to a lesser extent that stimulated by 10 PM isoproterenol and basal activity were also inhibited by these oligosaccharides. These data indicate that the inhibition of adenylyl cyclase activity by oligosaccharides was not specific to that stimulated by hCG and may affect the components of the adenylyl cyclase system directly.
u”
-3
P-AdrenergiLLgand,
Discussion
10
Glycopeptide or oligoskxharidq,
Since complete removal of carbohydrate gonadotropins and thyrotropin renders
from these
TABLE 1 DIRECT BINDING OF LABELED TO OVARY MEMBRANES
-9
OLIGOSACCHARIDES
Rat ovary membranes were incubated with ‘H-labeled oligosaccharides as described under Materials and Methods. The membranes were separated by centrifugation, solubilized and the radioactivity counted.
100
p.g/tube
Fig. 5. Effect
of glycopeptides and oligosaccharides on j% tor binding. Rat ovary membranes were in:aeJeegic.;. I-mdocyanopindolol ( - 150,000 cpm) in the presence of the indicated concentrations of &adrenergic ligands, glycopeptides or oligosacchatides in a total volume of 0.6 ml for 30 min at 32OC. Membrane bound l&and was separated as described under Materials and Methods and quantitated. Binding in the presence of 1 pg each of hCG and bTSH is also shown. Values are expressed as percent of control binding. (0) Propanolol; (0) isoproterenol; (0) at-acid glycoprotein glycopeptide; (M) ovalbumin glycopeptide; (A) hCG; (A) bTSH.
cpm bound hCG oligosaccharide 5 pl label ’ + 100 ng hCG + 100 pg ovalbumin glycopeptides + 7.8 pg hCG glycopeptides
117 127 151 149
Ovalbumin oligosacchari& 5 pl label b + 200 ng hCG + 12.5 pg hCG glycopeptides + 12.5 c(g transferrin glycopeptides + 12.5 /.tg ovalbumin glycopeptides
513 484 521 606 567
* 73,000 cpm; b 120,000 cpm of the oligosaccharide/tube.
hormones inactive (Goverman et al., 1982; Thotakura and Bahl, 1982; Bahl and Kalyan, 1983; Kalyan and Bahl, 1983; Amr et al., 1985) it is conceivable that the carbohydrate chain(s) of glycoprotein hormones maintain the hormone in a certain conformation required to activate the receptor. Alternatively, these chains may act in an as yet unknown manner with the hormone-receptor to bring about the required receptor conformation to cause coupling of the receptor to G protein and thus stimulate adenylyl cyclase activity. The
269
receptor-binding activity of the deglycosylated hormone remains unaffected or may actually be increased (Goverman et al., 1982; Bahl and Kalyan, 1983). We have previously demonstrated that the removal of carbohydrate affects the bioactivity of the hormone by impairing the coupling of the hormone occupied receptor to the G protein-adenylyl cyclase complex (Thotakura and Bahl, 1982). The exact mechanism by which the carbohydrate mediates the receptor-G protein coupling is not known. It has also been found that the addition of carbohydrate chains together with the deglycosylated hormone does not restore the hormone’s ability to stimulate adenylyl cyclase activity (unpublished). On the contrary, they inhibited adenylyl cyclase activity and also hCG binding to its receptor. In order to understand the role of carbohydrate at the molecular level we
have extended these observations using carbohydrate chains of different structures and studied their effect receptor binding and adenylyl cyclase stimulation of hCG. Calvo and Ryan (1985) have reported that glycopeptides specifically from hCG and hCGa, but not from fetuin or bovine y-globulins inhibited hCG-stimulated adenylyl cyclase but had no effect on the binding of hCG to receptor. They concluded that the glycopeptides inhibit the interaction of the carbohydrate chains of hCG with a lectin-like membrane component through which the hormone acts. We reasoned that if these effects were due to the carbohydrate part of the glycopeptides, oligosaccharides and glycopeptides from other proteins with similar carbohydrate structure would be expected to cause similar inhibition. It is also unlikely that a lectin would be
n q
Basal
hCG
ollgosaccharlde,
None
q +at
+Fet OLS
0
hCG
Acid GP OLS
B
+OvalbOLS
lsoprot
Forsk
pgltube
Fig. 6. A: Effect of hCG-oligosaccharide on hCG-stimulated adenylyl cyclase activity. Adenylyl cyclase activity was assayed in ovary membranes as described under Materials and Methods in the presence of 100 ng/tube of hCG and the indicated amounts of hCG-oligosaccharide. The total volume of the reaction mixture was 100 pl. Values represent the meanf SE of at least three determinations. B: Effect of oligosaccharides on hCG-stimulated adenylyl cyclase activity. Adenylyl cyclase activity was assayed in the presence of 100 ng/tube of hCG and 50 pg each of the oligosaccharides from different glycoproteins. The activity in the presence of hCG alone was taken as 100%. Values are averages of duplicate determinations which varied by less than 5%. (Isoprot) isoproterenol; (Forsk) forskolin; (al-Acid GP) al-acid glycoprotein; (Fet) fetuin; (Gvalb) ovalbmnin; (OLS) oligosaccharide.
270
specific for a glycan chain in its entirety. The present studies clearly support this contention. Glycopeptides as well as oligosaccharides prepared from hCG and several other functionally unrelated glycoproteins, transferrin, fetuin, al-acid glycoprotein and ovalbumin, all were able to inhibit the binding of hCG to its receptor on ovarian membranes. Similar inhibition was also observed using a purified soluble receptor preparation. Adenylyl cyclase activity stimulated by hCG was also inhibited by the oligosaccharides. The exact mechanism by which the free carbohydrate chains affect receptor binding of hCG and hCG-stimulated adenylyl cyclase activity is not clear. However, several inferences can be made from these results. First, since the inhibitory effect was caused by both glycopeptides and oligosaccharides, the carbohydrate moieties and not the peptides are responsible for this inhibition. An inhibition of hCG-receptor binding by the peptide portion of the glycopeptides can also be expected based on the observation that certain peptides from hCGa and hCG/3 were able to inhibit the receptor binding of hCG and bovine thyrotropin (Charlesworth et al., 1987; Keutman et al., 1988). As we have observed the inhibitory effect by oligosaccharides as well as by glycopeptides, it can be concluded that these effects are due to the carbohydrate chains. Second, the inhibition was not specific to only the carbohydrate chains from hCG, since it was also observed with bi-, tri- and tetra-antennary complex as well as high mannose type oligosaccharide units derived from other glycoproteins. These data suggest that common structural elements present in all these oligosaccharides are necessary to obtain the inhibitory effect. At present, the exact structural requirements of these common elements are unknown. Since all the monosaccharides that are constituents of the oligosaccharides, when individually tested in receptor binding assays were without effect, it involves more than a simple monosaccharide structure. Several synthetic oligosaccharides, which did not inhibit hCG-stimulated adenylyl cyclase activity in the previous report (Calvo and Ryan, 1985) presumably lack these required structural elements.
An interesting finding during the present studies was the ability of the glycopeptides to displace the hCG that was already bound to the receptor. Unlike non-polypeptide hormone receptors, e.g., adrenergic receptors, the dissociation rate of hCG from its receptor is very slow (Bruch et al., 1986). The oligosaccharide chains appear to increase the dissociation rate of receptor-occupied hCG. Thus, oligosaccharides may be used to elute hCG receptor from immobilized ligand during purification by affinity chromatography, since high salt and low pH which are generally used may be detrimental to the receptor activity. We propose that the observed effects of oligosaccharide chains on hCG-receptor interactions are due to the presence of a low affinity binding site for the carbohydrate moiety of the hormone (see below). At high concentrations, oligosaccharides or glycopeptides interact with this site, thereby inhibiting or displacing hCG from its peptide binding site on the receptor. This would also explain the lack of effect of the oligosaccharides on binding of ligands to j%adrenergic receptor, a non-glycoprotein hormone receptor. However, we were not successful in demonstrating any direct binding of labeled glycopeptides or oligosaccharides to ovary membranes, possibly due to the low affinity of the lectin-like site. In fact, it is difficult to separate the bound and free ligands with Kd in the millimolar to micromolar range even with the fastest separation techniques (Bennet and Yamamura, 1985). Alternatively, the oligosaccharides might affect the conformation of the hormone itself, in such a way that it no longer can bind to the receptor. In any case, the carbohydrate effects appear to be specific to the ovary, since glucagon-stimulated adenylyl cyclase activity of liver membranes was not affected by hCG glycopeptides (Calvo and Ryan, 1985). Our proposal of a lectin-like site on the receptor itself and not on another membrane protein as previously proposed (Calvo and Ryan, 1985), stems from two lines of evidence. First, the oligosaccharide effects on hCG-receptor interactions are observed with a pure receptor preparation as well as with the membrane receptor. Second, a recent observation of homology between a stretch of peptide chain on cloned hCG receptor and soybean
271
lectin lends support to the above contention (McFarland et al., 1989). More recently, a lectin domain has been revealed in cloned lymphocyte homing receptor (La&y et al., 1989). In the present study, glycopeptides and oligosaccharides inhibited receptor-binding and adenylyl cyclase-stimulating activity of hCG at high concentrations, in the pg/rnl range. This is parallel to the effects of synthetic peptides or the subunits of hCG and other glycopeptide hormones. The cu-subunit of hCG or synthetic peptides of hCGa and -8 were shown to bind to and displace labeled hCG from hCG receptor with low affinity (Charlesworth et al., 1987; Keutman et al., 1988). Synthetic peptides from hCGa were also shown to inhibit TSH-stimulated cyclic AMP production, although these peptides themselves do not possess any activity (Morris et al., 1988). However, the effects of glycopeptides observed at such high concentrations may not be of any physiological significance. In general, the carbohydrate moieties appear to be involved in the binding of the hormone to its receptor. Removal of sialic acid or most of the glycan moieties actually increases the affinity of the hormone to its receptor (Moyle et al., 1975; Kalyan and Bahl, 1983). It has also been shown that abnormal hCG molecules with multiantennary carbohydrate chains possess very low affinity to the receptor and low biological activity (Nishimura et al., 1981). Carbohydrate chains of glycoprotein hormones thus appear to modulate the binding affinity of the hormones to their receptors, although it is not established if the carbohydrate moieties of the receptor itself are involved. The observed inhibition of adenylyl cyclase activity of ovary membranes, basal as well as that stimulated by hCG, isoproterenol and forskolin suggests that the carbohydrate effect might be directly on the enzyme itself, rather than specific to that stimulated by hCG. In conclusion, our results suggest that the effects of Asn-linked oligosaccharide chains on hCG-stimulated adenylyl cyclase are not limited to glycopeptides from hCG and hCGa, but are observed with glycopeptides and oligosaccharides from various other glycoproteins. In addition, they inhibit hCG binding to the membrane receptor and to the purified soluble receptor. These effects
may be related to a putative lectin-like site on the receptor itself and not a separate membrane component.
Ann, S., Menezes-Ferreira, M., Shimohigashi, Y., Chen, H.C., Nisula, B.C. and Weintraub, B.D. (1985) J. Endocrinol. Invest. 8, 537-541. BahI, O.P. and Kalyan, N.K. (1983) in Role of Peptides and Proteins in Control of Reproduction (McCann, S.M. and Dhindsa, D.S., eds.), pp. 293-327, Elsevier, New York. BahI, O.P., Thotakura, N.R. and Anumula, K.R. (1984) Biological role of carbohydrates in gonadotropins, in Hormone Receptors in Growth and Reproduction (Saxena, B.B. et al., eds.), Serono Vol. 9, Chapter 11, Raven Press, New York. Bennett, Jr., J.P. and Yamamura, HI. (1985) Neurotransmitter, hormone, or drug receptor binding methods, in Neurotransmitter Receptor Binding (Yamamura, HI., Enna, S.J. and Kuhar, M.J., eds.), pp. 61-89, Raven Press, New York. Bruch, R.C., Thotakura, N.R. and BahI, O.P. (1986) J. Biol. Chem. 261, 9450-9460. CaIvo, F.O. and Ryan, R.J. (1985) Biochemistry 24,1953-1959. Charlesworth, M.C., McCormick, D.J., Madden, B. and Ryan, R.J. (1987) J. Biol. Chem. 262, 13409-13416. Goverman, J.M., Parsons, T.F. and Pierce, J.G. (1982) J. Biol. Chem. 257, 15059-15064. Kalyan, N.K. and BahI, O.P. (1983) J. Biol. Chem. 258, 67-74. Kessler, M.J., Reddy, M.S., Shah, R.H. and BahI, O.P. (1979) J. Biol. Chem. 254, 7901-7908. Keutman, H.T., Charlesworth, M.C., Kitzmann, K., Mason, K.A., Johnson, L. and Ryan, R.J. (1988) Biochemistry 27, 8939-8944. Kobata, A. (1984) in Biology of Carbohydrates (Ginsburg, V. and Robbins, P.W., eds.), Vol. 2, Chapter 2, John Wiley and Sons, New York. Lasky, L.A., Singer, M.S., Yednock, T.A., Dowbenko, D., Fennie, C., Rodriguez, H., Nguyen, T., Stachel, S. and Rosen, S.D. (1989) Ceil 56, 1045-1055. Manjunath, P. and Sairam, M.R. (1982) J. Biol. Chem. 257, 7109-7115. McFarland, K.C., Sprenghl, R., Philhps, H.S., KohIer, M., Rosembht, N., NikoIics, K., Segaloff, D.L. and Seeburg, P.H. (1989) Science 245,494-499. Morris, III, J.C., Jiang, N.S., Charlesworth, M.C., McCormick, D.J. and Ryan, R.J. (1988) Endocrinology 123, 456-462. Moyle, W.R., BahI, O.P. and Marz, L. (1975) J. Biol. Chem. 250, 9163-9169. Nishirnura, R., Endo, Y., Tanabe, K., Ashitaka, Y. and Tojo, S. (1981) J. Endocrinol. Invest. 4, 349-355. Pierce, J.G. and Parsons, T.F. (1981) Annu. Rev. B&hem. 50, 465-495. Rademacher, T.W., Parekh, R.B. and Dwek, R.A. (1988) Annu. Rev. Biochem. 57, 785-838. Ryan, R.J., Keutman, H.T., Charlesworth, M.C., McCormick,
272 D.J., Milius, R.P., Calve, F.O. and Vutyavanich, T. (1987) Recent Rag. Harm. Res. 43, 383-429. Salomon, Y. (1979) Adv. Cyclic Nucleotide Res. 10, 35-55. Takasaki, S., Mizuochi, T. and Kobata, A. (1982) Methods Enzymol. 83, 263-268.
Thotakura, N.R. and Bahl, O.P. (1982) B&hem. Biophys. Res. Commun. 108, 399-405. Yamaahita, K., Mizuochi, T. and Kobata, A. (1982) Methods Enzymol. 83, 105-126.
Molecular and Cellular Enakcrinology, 70 (1990) 263-272 Elsevier Scientific Publishers Ireland, Ltd.
MOLCEL 02279
The role of carb.ohydrate in human choriogonadotropin
(hCG) action
Effects of N-linked carbohydrate chains from hCG and other glycoproteins on hormonal activity N. Rao Thotakura ‘, Bruce D. Weintraub
’
and Om P. Bahl 2
’ Molecular, Cellular and Nutritional Endocrinology Branch, NIDDK NIH, Bethesda, MD, U.S.A., and * Department of Biological Sciences, State Universiry of New York at Buffalo, Amherst, NY, U.S.A. (Received 11 January 1990; accepted 7 February 1990)
Key words: Glycoprotein; Gonadotropin;
Carbohydrate chain; Receptor binding; Hormone; N-linked oligosaccharide;_(Rat
ovary)
Deglycosylation of gonadotropins and thyrotropin results in a major loss of hormonal bioactivity, while not impairing receptor-binding activity. However, a direct role of the glycan moieties in hormonal signal transduction has not been demonstrated. The addition of carbohydrate chains together with the deglycosylated hormone does not restore the hormonal activity. In contrast, glycopeptides were found to inhibit human choriogonadotropin (hCG)-stimulated adenylyl cyclase activity and hCG binding to its receptor. An inhibition of hCG-stimulated adenylyl cyclase activity but not hCG binding to receptor by glycopeptides specifically from hCG, has previously been reported as a lectin-like membrane component has been implicated in hCG action. In the present study we have shown that glycopeptides and oligosaccharides prepared from hCG, transferrin, fetuin, q-acid glycoprotein and ovalbumin inhibit the binding of hCG to its receptor. The inhibition was also observed with a highly purified preparation of the receptor, thus suggesting a lack of involvement of other lectin-like membrane components as previously proposed. We suggest that a lectin-like interaction with the hormone, if any, involves the receptor itself. Adenylyl cyclase activity stimulated by hCG, isoproterenol or forskolin was inhibited by oligosaccharides, indicating a non-specific interaction. Our results suggest that Asn-linked oligosaccharide chains from various glycoproteins perturb hCG-receptor interactions through a putative carbohydrate binding site on the receptor.
Introduction
The glycoprotein hormones, choriogonadotropin (CG), lutropin, follitropin and thyrotropin are
Address for correspondence: N. Rae Thotakura, Building 10, Room 8D14, NIH, Bethesda, MD 20892, U.S.A. Supported in part by USPHS grant HD 08766.
composed of two dissimilar, non-covalently linked subunits, cu and 8. The common a-subunit of glycoprotein hormones contains two Asn-linked carbohydrate chains, while the /&subunit contains one or two Asn-linked units depending upon the hormone (Pierce and Parsons, 1981). The majority of these carbohydrate chains are biantermary complex type with minor populations of mono- and triantennary chains (Kessler et al., 1979; Kobata, 1984).
264
The importance of carbohydrates in the function of glycoprotein hormones has been well documented (Manjunath and Sairam, 1982; Bahl et al., 1984; Ryan et al., 1987). Enzymatic (Moyle et al., 1975; Goverman et al., 1982) or chemical deglycosylation of hCG and other glycoprotein hormones results in a loss of their post-receptor biological activity (Thotakura and Bahl, 1982; Bahl and Kalyan, 1983; Kalyan and Bahl, 1983). Data from our laboratory have shown that incubation of ovarian membrane hCG receptor with deglycosylated hCG and glycopeptides from hCG, cri-acid glycoprotein or ovalbumin did not restore the adenylyl cyclase-stimulating activity of the hormone (unpublished). The exact mechanism of how the carbohydrate chains on hCG bring about the activation of the hormone-occupied receptor is not known. It is currently believed (Kobata, 1984; Rademacher et al., 1988) that the carbohydrate chains of hCG bind to a lectin-like membrane component and thereby mediate the activation of adenylyl cyclase. These conclusions stemmed from a report (Calvo and Ryan, 1985) showing an inhibition of hCG-stimulated adenylyl cyclase but not receptor binding of HCG by glycopeptides specifically from hCG. If this effect is due to the oligosaccharides, it is hard to conceive that hCG glycopeptides show specific effects, since several other glycoproteins contain structurahy similar oligosaccharide chains identical to those on hCG. Moreover, in our studies, we have found inhibition of adenylyl cyclase as well as receptor binding of hCG by glycopeptides from several other glycoproteins. Since the role of carbohydrate in glycoprotein hormone action is an important question, we have reevaluated the effects of Asn-linked carbohydrate chains on hCG action using glycopeptides as well as oligosaccharides from hCG and various other functionally unrelated glycoproteins like transferrin, fetuin, cu,-acid glycoprotein and ovalbumin containing different types (bi-, triand tetra-antemrary as well as high marmose type) of carbohydrate chains. Materials and methods Oligosaccharides
and glycopeptides
Oligosaccharides from hCG, transferrin, fetuin, ovalbumin and cy,-acid glycoprotein were pre-
pared by hydrazinolysis according to the procedure of Takasaki et al. (1982). Lyophilized samples of the glycoproteins were hydrolyzed with anhydrous hydrazine for lo-12 h at 110°C in a sealed tube. After drying and removing the hydrazine, the samples were dissolved in saturated NaHCO, and N-acetylated by the addition of acetic anhydride. The oligosaccharides were purified on a Dowex 50 X 8-H+ column and then on a Bio Gel P-4 column. An aliquot of each of the oligosaccharides was reduced with NaB3H, in 0.05 N NaOH for 4 h at 30’ C. The samples were then neutralized with 1 N CH,COOH, purified by Dowex chromatography and the borohydride removed by repeated evaporation with methanol. After N-acetylation, the oligosaccharides were purified on a Bio Gel P-4 column (Yamashita et al., 1982). To assess their purity, an aliquot of each of the oligosaccharides was reduced with NaB3H, (Takasaki et al., 1982), chromatographed on Silica Gel 60 (Merck) in n-propanol/acetic acid/water (3 : 3 : 2) and autoradiographed. All the oligosaccharides prepared showed a single spot on thin-layer chromatography in the above system. The structures of all these oligosaccharides have previously been established (Kobata, 1984). Glycopeptides from hCG, ovalbumin and (piacid glycoprotein were prepared by extensive pronase digestion as described (Kessler et al., 1979). Briefly, the glycoproteins were reduced with dithiothreitol and the disulfide groups were carboxamidomethylated with iodoacetamide. The glycoprotein samples were digested with pronase for 96 h at 37’ C in an atmosphere of nitrogen with 1% by weight of the enzyme added at 24 h intervals. The resulting peptides were separated on Sephadex G-SO(fine) and the fraction were monitored for carbohydrate using phenol-sulfuric acid reaction. The glycopeptide fractions were pooled, desalted and lyophilized. Receptor binding assays
Receptor binding assays for hCG were performed using superovulated rat ovary membranes (Thotakura and Bahl, 1982). Ovaries were homogenized in 50 mM Tris-HCI, pH 7.2 containing 1 mg/ml bovine serum albumin and centrifuged at 10,000 X g for 20 min. The membrane pellet was resuspended in the same buffer and incubated
265
with 100,000 cpm 1251-hCG (35-50 $i/pg, iodinated by the chloramine T method) for 1 h at 37 o C in the presence or absence of various concentrations of unlabeled hCG and/or glycopeptides and oligosaccharides. Bound label was counted after separating the membranes by centrifugation for 15 min at 1500 X g. In experiments with pure receptor, an aliquot of affinitypurified, homogenous soluble receptor preparation was incubated in 50 mM Tris-HCl, pH 7.4 containing 20% glycerol, with labeled hCG for 18 h at 4 o C, and the bound label was separated by precipitation with poly(ethylene glycol) 8000, at a final concentration of 10% (Bruch et al., 1986). Beta-adrenergic receptor binding was assayed by incubat~g membranes prepared as above with ‘251-labeled iodocyanopindolol (ICYP) and the effectors in an incubation buffer containing 25 mM Tris-HCl, pH 7.4,l mM EDTA, 1 mg/ml bovine serum albumin, 0.1 mM ascorbic acid, 5 mM MgCl, for 30 min at 32” C. The membranes were collected by quickly filtering on GF/C filters under vacuum, washed twice with buffer and counted for bound radioactivity. In experiments testing the direct binding of carbohydrate chains to the membranes, two ovaries where homogenized in 10 ml of 50 mM Tris-HCl, pH 7.2 containing 1 mg/ml bovine serum albumin. Oligosaccharides labeled as described above, in 0.2 ml were incubated with 0.5 ml ovary membranes in 50 mM Tris-HCl, pH 7.2 containing 1 mg/ml bovine serum albumin for 1 h at 37” C. After separation of the membranes by centrifugation, the pellet was solubilized and counted in a liquid scintillation counter. Adenylyl cyclase assay Ovary membranes as prepared
above were incubated with hCG and effecters in an assay mixture containing 50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mg/ml bovine serum albumin, 2 mM ATP, 5 mM MgCl,, 10 PM GPP(NH)P, 0.2 mM SQ20009, 10 mM creatine phosphate, 50 U/ml tireanne kinase and 0.5 &i of [o-32P]ATP in a total volume of 0.1 ml. Incubations were carried out for 15 min at 34OC and terminated by the addition of excess unlabeled ATP, about 10,000 cpm of ‘H-cyclic AMP and 0.1% sodium dodecyl sulfate. The cyclic AMP formed was determined
by sequential chromatography on Dowex alumina columns (SaIomon, 1979).
and
Resuits Effect of glycopeptides and oligosaccharides on the binding of hCG to its receptor
We have previously observed that incubation of ovary membranes with deglycosylated hCG and glycopeptides from hCG, ovalbumin or or-acid glycoprotein failed to restore the adenylyl cyclase-stimulating activity of the hormone. However, the glycopeptides were found to inhibit adenylyl cyclase activity stimulated by hCG and further, in receptor binding studies, the glycopeptides were found to inhibit hCG binding to its receptor. As shown in Fig. 1, both glycopeptides and oligosaccharides, not only from hCG, but from a,-acid glycoprotein, ovalbumin and transferrin inhibited ‘*%hCG binding to its receptor on superovulated rat ovary membranes, This effect was observed at very high ~n~ntrations of the glycopeptides and oligosaccharides, three to four orders of magnitude higher than that of hCG required to cause similar displacement. The difference in the potencies of the various glycopeptides and o~gosa~h~des is probably related to their different carbohydrate contents. Preincubation of the ovary membranes with the glycopeptides also showed identical inhibition curves (not shown). Since the inhibition was noted with both glycopeptides as well as oligosaccharides, and peptide fractions with no ~bohy~ate chains are without any effect (not shown), we conclude that the observed effects are due to the carbohydrate moieties and not the peptide component of the glycopeptides. Since the inhibition of hCG binding was caused by the ~bohydrate, further experiments were performed to determine if monosaccharides could cause the same effects. Galactose, marmose and N-acetyl glucosamine at concentrations up to 1 mg and sialic acids up to 250 pg per assay did not have any effect on the binding of hCG to its receptor in ovary membrane receptor binding assay (Fig. 2). These data indicate that for the inhibitory effect of carbohydrates on hCG binding, a simple monosaccharide structure is not sufficient and rather, some common structural fea-
266
-
Further studies were performed to validate the involvement of any lectin-like membrane component in the inhibition of hCG to its receptor by oligosaccharides, as previously proposed (Calvo and Ryan, 1985). We have previously purified hCG/lutropin receptor from rat ovary (Bruch et al., 1986), and using this preparation, binding studies were performed in the presence of glycopeptides and oligosaccharides. Fig. 4 shows that the glycopeptides and oligosaccharides inhibited hCG binding to pure receptor as they had in membrane receptor assays. This shows that the effect of glycopeptides and oligosaccharides on hCG binding to its receptor does not require the integrity of membrane and does not involve any other membrane component.
L,.
0 0
“>..
..
““.“n
..I
10 hCG,
Glycopeptide
or
1
120
-i
100
1000
ng/tube
ollgosaccharlda,
ugitube
Fig. 1. Effect of glycopeptides and oligosaccharides on hCG binding to membrane receptor. Bat ovary membranes were incubated with ‘251-hCG ( -100,000 cpm) and the indicated concentrations of unlabeled hCG or various glycopeptides and oligosaccharides in 50 mM Tris-HCl containing 1 mg/ml bovine serum albumin in a total volume of 1 ml for 1 h at 37OC. Membranes were separated by centrifugation and the bound label quantitated. Binding in the absence of any effector was usually about 20% of the total counts added and is taken as 100%. The data shown are from a typical experiment re peated at least 3 times. (+) hCG; (A) cx,-acid glycoprotem glycopeptide; (W) ovalbumin glycopeptide; (0) hCG glycopeptide; (0) hCG oligosaccharide; (A) transferrin oligosaccharide; (0) ovalbumin oligosaccharide.
100
q q 0
Control GA Man GlcNAc Slalic Acids $“:3”$gp
GP
I
s E
80
z *0 2 i
60
z I : 40
ture(s) present in high mannose and complex type oligosaccharides is required. Another interesting observation was the ability of the glycopeptides to displace the hormone already bound to the receptor. The binding of hCG to its receptor is of high affinity and is reversible very slowly over a period of several hours as has been demonstrated previously (Bruch et al., 1986). However, when hCG was allowed to achieve equilibrium binding by incubation with the receptor for 1 h at 37” C, and further incubated with glycopeptides at concentrations comparable to those used in Fig. 1, a very rapid reversal of the hormone binding was observed (Fig. 3). Again, monosaccharides were without effect (not shown).
20
0
Fig. 2. Effect of monosaccharides on hCG binding to receptor. Bat ovary membranes were incubated with “‘1-hCG and the indicated monosaccharides under conditions as described in Fig. 1. Gal, Man and GlcNAc were each present at 1 mg/tube and sialic acids at 250 pg. Inhibition in the presence of hCG and glycopeptides is also shown for comparison. cut-Acid glycoprotein glycopeptide (cd Acid gp GP) and ovalbumin glycopeptide (Oval GP) were added at 500 pg each per tube. Values are mean f SE of triplicate determinations.
261
w GL.P ILP
25
and its receptor or to a non-specific effect on membrane proteins, we took advantage of the presence of /3-adrenergic receptors in the oy2y (Thotakura and Bahl, 1982). Binding of Icyanopindolol to ovary membranes was measured in the presence of different concentrations of ovalbumin oligosaccharides and q-acid glycoprotein glycopeptides. As shown in Fig. 5, the oligosaccharides and glycopeptides did not affect j?adrenergic receptor binding in the same membrane preparations in which hCG binding was inhibited. Moreover, hCG and bovine thyrotropin at 1 pg/O.S ml concentration were without any effect. These results indicate that the effect of the carbohydrate units is specific to hCG binding to its receptor.
0
Fig. 3. Effect of glycopeptides on the dissociation of receptorbound hCG. Ovary membranes were incubated with 12?-hCG for 1 h during the first incubation and the incubation was continued for an additional 1 h after adding 500 cg each of q-acid glycoprotein glycopeptide (al Acid gp GLP) and ovalbumin glycopeptide (Gvalb GLP), 1 pg hCG or none in the control. The amount of label bound in control is taken as 100%. Values are mean f SE of triplicate determinations.
One possible explanation for the inhibition of hCG-receptor binding by oligosaccharides is that the receptor has a low affinity binding site for the carbohydrate, which is separate from the polypeptide binding site and the binding of glycopeptides to this low affinity site inhibits the binding of hCG-carbohydrate, and thereby displaces hCG from its receptor. To test this hypothesis, direct binding of 3H-labeled oligosaccharides from hCG and ovalbumin to ovarian receptor was studied. We were unable to demonstrate any significant specific binding of the labeled oligosaccharides to ovary membranes as shown in Table 1. Similar results were obtained when glycopeptides labeled with 12? using Bolton-Hunter reagent were used (not shown). Effect of glycopeptides and oligosaccharides on j3adrenergic receptor binding To determine whether the effect of glycopeptides and oligosaccharides was specific to hCG
OF““‘i
.“‘lrl hCG,
Glycopeptide
or
“’
’
100
l(
IO
ngltube
ollgosaccharlde,
vgmba
Fig. 4. Effect of glycopeptides and oligosaccharides on hCG binding to purified soluble receptor. Affinity-purified soluble ovarian hCG receptor was incubated with the indicated concentrations of unlabeled hCG or glycopeptides or oligosaccharides for 18 h at 4’ C. Receptor-bound label was separated by precipitating with poly(ethylene glycol) as described under Materials and Methods. Values are plotted as percent of label bound in the absence of any effector. Data shown are from a representative experiment repeated at least 3 times. (+) hCG; (0) q-acid glycoprotein oligosaccharide; (0) fetuin oligosaccharide; (A) q-acid glycoprotein glycopeptide; (m) ovalbumin glycopcptide.
268
Effect of oligosaccharides ity
on adenylyl cyclase activ-
To determine whether the inhibitory effects of the glycopeptides and oligosaccharides on receptor binding can be observed at the post-receptor actions of the hormone, we have assayed adenylyl cyclase activity of ovary membranes stimulated by hCG, in the presence of various preparations of oligosaccharides. Fig. 6A shows a dose-dependent inhibition of adenylyl cyclase activity stimulated by 100 ng hCG by oligosaccharide prepared from hCG. Further studies showed (Fig. 6B) that this effect was not specific to hCG-oligosaccharides. Oligosaccharides from fetuin, a,-acid glycoprotein and ovalbumin were also able to inhibit hCGstimulated adenylyl cyclase activity. Moreover, the activity stimulated by 10 PM forskolin and to a lesser extent that stimulated by 10 PM isoproterenol and basal activity were also inhibited by these oligosaccharides. These data indicate that the inhibition of adenylyl cyclase activity by oligosaccharides was not specific to that stimulated by hCG and may affect the components of the adenylyl cyclase system directly.
u”
-3
P-AdrenergiLLgand,
Discussion
10
Glycopeptide or oligoskxharidq,
Since complete removal of carbohydrate gonadotropins and thyrotropin renders
from these
TABLE 1 DIRECT BINDING OF LABELED TO OVARY MEMBRANES
-9
OLIGOSACCHARIDES
Rat ovary membranes were incubated with ‘H-labeled oligosaccharides as described under Materials and Methods. The membranes were separated by centrifugation, solubilized and the radioactivity counted.
100
p.g/tube
Fig. 5. Effect
of glycopeptides and oligosaccharides on j% tor binding. Rat ovary membranes were in:aeJeegic.;. I-mdocyanopindolol ( - 150,000 cpm) in the presence of the indicated concentrations of &adrenergic ligands, glycopeptides or oligosacchatides in a total volume of 0.6 ml for 30 min at 32OC. Membrane bound l&and was separated as described under Materials and Methods and quantitated. Binding in the presence of 1 pg each of hCG and bTSH is also shown. Values are expressed as percent of control binding. (0) Propanolol; (0) isoproterenol; (0) at-acid glycoprotein glycopeptide; (M) ovalbumin glycopeptide; (A) hCG; (A) bTSH.
cpm bound hCG oligosaccharide 5 pl label ’ + 100 ng hCG + 100 pg ovalbumin glycopeptides + 7.8 pg hCG glycopeptides
117 127 151 149
Ovalbumin oligosacchari& 5 pl label b + 200 ng hCG + 12.5 pg hCG glycopeptides + 12.5 c(g transferrin glycopeptides + 12.5 /.tg ovalbumin glycopeptides
513 484 521 606 567
* 73,000 cpm; b 120,000 cpm of the oligosaccharide/tube.
hormones inactive (Goverman et al., 1982; Thotakura and Bahl, 1982; Bahl and Kalyan, 1983; Kalyan and Bahl, 1983; Amr et al., 1985) it is conceivable that the carbohydrate chain(s) of glycoprotein hormones maintain the hormone in a certain conformation required to activate the receptor. Alternatively, these chains may act in an as yet unknown manner with the hormone-receptor to bring about the required receptor conformation to cause coupling of the receptor to G protein and thus stimulate adenylyl cyclase activity. The
269
receptor-binding activity of the deglycosylated hormone remains unaffected or may actually be increased (Goverman et al., 1982; Bahl and Kalyan, 1983). We have previously demonstrated that the removal of carbohydrate affects the bioactivity of the hormone by impairing the coupling of the hormone occupied receptor to the G protein-adenylyl cyclase complex (Thotakura and Bahl, 1982). The exact mechanism by which the carbohydrate mediates the receptor-G protein coupling is not known. It has also been found that the addition of carbohydrate chains together with the deglycosylated hormone does not restore the hormone’s ability to stimulate adenylyl cyclase activity (unpublished). On the contrary, they inhibited adenylyl cyclase activity and also hCG binding to its receptor. In order to understand the role of carbohydrate at the molecular level we
have extended these observations using carbohydrate chains of different structures and studied their effect receptor binding and adenylyl cyclase stimulation of hCG. Calvo and Ryan (1985) have reported that glycopeptides specifically from hCG and hCGa, but not from fetuin or bovine y-globulins inhibited hCG-stimulated adenylyl cyclase but had no effect on the binding of hCG to receptor. They concluded that the glycopeptides inhibit the interaction of the carbohydrate chains of hCG with a lectin-like membrane component through which the hormone acts. We reasoned that if these effects were due to the carbohydrate part of the glycopeptides, oligosaccharides and glycopeptides from other proteins with similar carbohydrate structure would be expected to cause similar inhibition. It is also unlikely that a lectin would be
n q
Basal
hCG
ollgosaccharlde,
None
q +at
+Fet OLS
0
hCG
Acid GP OLS
B
+OvalbOLS
lsoprot
Forsk
pgltube
Fig. 6. A: Effect of hCG-oligosaccharide on hCG-stimulated adenylyl cyclase activity. Adenylyl cyclase activity was assayed in ovary membranes as described under Materials and Methods in the presence of 100 ng/tube of hCG and the indicated amounts of hCG-oligosaccharide. The total volume of the reaction mixture was 100 pl. Values represent the meanf SE of at least three determinations. B: Effect of oligosaccharides on hCG-stimulated adenylyl cyclase activity. Adenylyl cyclase activity was assayed in the presence of 100 ng/tube of hCG and 50 pg each of the oligosaccharides from different glycoproteins. The activity in the presence of hCG alone was taken as 100%. Values are averages of duplicate determinations which varied by less than 5%. (Isoprot) isoproterenol; (Forsk) forskolin; (al-Acid GP) al-acid glycoprotein; (Fet) fetuin; (Gvalb) ovalbmnin; (OLS) oligosaccharide.
270
specific for a glycan chain in its entirety. The present studies clearly support this contention. Glycopeptides as well as oligosaccharides prepared from hCG and several other functionally unrelated glycoproteins, transferrin, fetuin, al-acid glycoprotein and ovalbumin, all were able to inhibit the binding of hCG to its receptor on ovarian membranes. Similar inhibition was also observed using a purified soluble receptor preparation. Adenylyl cyclase activity stimulated by hCG was also inhibited by the oligosaccharides. The exact mechanism by which the free carbohydrate chains affect receptor binding of hCG and hCG-stimulated adenylyl cyclase activity is not clear. However, several inferences can be made from these results. First, since the inhibitory effect was caused by both glycopeptides and oligosaccharides, the carbohydrate moieties and not the peptides are responsible for this inhibition. An inhibition of hCG-receptor binding by the peptide portion of the glycopeptides can also be expected based on the observation that certain peptides from hCGa and hCG/3 were able to inhibit the receptor binding of hCG and bovine thyrotropin (Charlesworth et al., 1987; Keutman et al., 1988). As we have observed the inhibitory effect by oligosaccharides as well as by glycopeptides, it can be concluded that these effects are due to the carbohydrate chains. Second, the inhibition was not specific to only the carbohydrate chains from hCG, since it was also observed with bi-, tri- and tetra-antennary complex as well as high mannose type oligosaccharide units derived from other glycoproteins. These data suggest that common structural elements present in all these oligosaccharides are necessary to obtain the inhibitory effect. At present, the exact structural requirements of these common elements are unknown. Since all the monosaccharides that are constituents of the oligosaccharides, when individually tested in receptor binding assays were without effect, it involves more than a simple monosaccharide structure. Several synthetic oligosaccharides, which did not inhibit hCG-stimulated adenylyl cyclase activity in the previous report (Calvo and Ryan, 1985) presumably lack these required structural elements.
An interesting finding during the present studies was the ability of the glycopeptides to displace the hCG that was already bound to the receptor. Unlike non-polypeptide hormone receptors, e.g., adrenergic receptors, the dissociation rate of hCG from its receptor is very slow (Bruch et al., 1986). The oligosaccharide chains appear to increase the dissociation rate of receptor-occupied hCG. Thus, oligosaccharides may be used to elute hCG receptor from immobilized ligand during purification by affinity chromatography, since high salt and low pH which are generally used may be detrimental to the receptor activity. We propose that the observed effects of oligosaccharide chains on hCG-receptor interactions are due to the presence of a low affinity binding site for the carbohydrate moiety of the hormone (see below). At high concentrations, oligosaccharides or glycopeptides interact with this site, thereby inhibiting or displacing hCG from its peptide binding site on the receptor. This would also explain the lack of effect of the oligosaccharides on binding of ligands to j%adrenergic receptor, a non-glycoprotein hormone receptor. However, we were not successful in demonstrating any direct binding of labeled glycopeptides or oligosaccharides to ovary membranes, possibly due to the low affinity of the lectin-like site. In fact, it is difficult to separate the bound and free ligands with Kd in the millimolar to micromolar range even with the fastest separation techniques (Bennet and Yamamura, 1985). Alternatively, the oligosaccharides might affect the conformation of the hormone itself, in such a way that it no longer can bind to the receptor. In any case, the carbohydrate effects appear to be specific to the ovary, since glucagon-stimulated adenylyl cyclase activity of liver membranes was not affected by hCG glycopeptides (Calvo and Ryan, 1985). Our proposal of a lectin-like site on the receptor itself and not on another membrane protein as previously proposed (Calvo and Ryan, 1985), stems from two lines of evidence. First, the oligosaccharide effects on hCG-receptor interactions are observed with a pure receptor preparation as well as with the membrane receptor. Second, a recent observation of homology between a stretch of peptide chain on cloned hCG receptor and soybean
271
lectin lends support to the above contention (McFarland et al., 1989). More recently, a lectin domain has been revealed in cloned lymphocyte homing receptor (La&y et al., 1989). In the present study, glycopeptides and oligosaccharides inhibited receptor-binding and adenylyl cyclase-stimulating activity of hCG at high concentrations, in the pg/rnl range. This is parallel to the effects of synthetic peptides or the subunits of hCG and other glycopeptide hormones. The cu-subunit of hCG or synthetic peptides of hCGa and -8 were shown to bind to and displace labeled hCG from hCG receptor with low affinity (Charlesworth et al., 1987; Keutman et al., 1988). Synthetic peptides from hCGa were also shown to inhibit TSH-stimulated cyclic AMP production, although these peptides themselves do not possess any activity (Morris et al., 1988). However, the effects of glycopeptides observed at such high concentrations may not be of any physiological significance. In general, the carbohydrate moieties appear to be involved in the binding of the hormone to its receptor. Removal of sialic acid or most of the glycan moieties actually increases the affinity of the hormone to its receptor (Moyle et al., 1975; Kalyan and Bahl, 1983). It has also been shown that abnormal hCG molecules with multiantennary carbohydrate chains possess very low affinity to the receptor and low biological activity (Nishimura et al., 1981). Carbohydrate chains of glycoprotein hormones thus appear to modulate the binding affinity of the hormones to their receptors, although it is not established if the carbohydrate moieties of the receptor itself are involved. The observed inhibition of adenylyl cyclase activity of ovary membranes, basal as well as that stimulated by hCG, isoproterenol and forskolin suggests that the carbohydrate effect might be directly on the enzyme itself, rather than specific to that stimulated by hCG. In conclusion, our results suggest that the effects of Asn-linked oligosaccharide chains on hCG-stimulated adenylyl cyclase are not limited to glycopeptides from hCG and hCGa, but are observed with glycopeptides and oligosaccharides from various other glycoproteins. In addition, they inhibit hCG binding to the membrane receptor and to the purified soluble receptor. These effects
may be related to a putative lectin-like site on the receptor itself and not a separate membrane component.
Ann, S., Menezes-Ferreira, M., Shimohigashi, Y., Chen, H.C., Nisula, B.C. and Weintraub, B.D. (1985) J. Endocrinol. Invest. 8, 537-541. BahI, O.P. and Kalyan, N.K. (1983) in Role of Peptides and Proteins in Control of Reproduction (McCann, S.M. and Dhindsa, D.S., eds.), pp. 293-327, Elsevier, New York. BahI, O.P., Thotakura, N.R. and Anumula, K.R. (1984) Biological role of carbohydrates in gonadotropins, in Hormone Receptors in Growth and Reproduction (Saxena, B.B. et al., eds.), Serono Vol. 9, Chapter 11, Raven Press, New York. Bennett, Jr., J.P. and Yamamura, HI. (1985) Neurotransmitter, hormone, or drug receptor binding methods, in Neurotransmitter Receptor Binding (Yamamura, HI., Enna, S.J. and Kuhar, M.J., eds.), pp. 61-89, Raven Press, New York. Bruch, R.C., Thotakura, N.R. and BahI, O.P. (1986) J. Biol. Chem. 261, 9450-9460. CaIvo, F.O. and Ryan, R.J. (1985) Biochemistry 24,1953-1959. Charlesworth, M.C., McCormick, D.J., Madden, B. and Ryan, R.J. (1987) J. Biol. Chem. 262, 13409-13416. Goverman, J.M., Parsons, T.F. and Pierce, J.G. (1982) J. Biol. Chem. 257, 15059-15064. Kalyan, N.K. and BahI, O.P. (1983) J. Biol. Chem. 258, 67-74. Kessler, M.J., Reddy, M.S., Shah, R.H. and BahI, O.P. (1979) J. Biol. Chem. 254, 7901-7908. Keutman, H.T., Charlesworth, M.C., Kitzmann, K., Mason, K.A., Johnson, L. and Ryan, R.J. (1988) Biochemistry 27, 8939-8944. Kobata, A. (1984) in Biology of Carbohydrates (Ginsburg, V. and Robbins, P.W., eds.), Vol. 2, Chapter 2, John Wiley and Sons, New York. Lasky, L.A., Singer, M.S., Yednock, T.A., Dowbenko, D., Fennie, C., Rodriguez, H., Nguyen, T., Stachel, S. and Rosen, S.D. (1989) Ceil 56, 1045-1055. Manjunath, P. and Sairam, M.R. (1982) J. Biol. Chem. 257, 7109-7115. McFarland, K.C., Sprenghl, R., Philhps, H.S., KohIer, M., Rosembht, N., NikoIics, K., Segaloff, D.L. and Seeburg, P.H. (1989) Science 245,494-499. Morris, III, J.C., Jiang, N.S., Charlesworth, M.C., McCormick, D.J. and Ryan, R.J. (1988) Endocrinology 123, 456-462. Moyle, W.R., BahI, O.P. and Marz, L. (1975) J. Biol. Chem. 250, 9163-9169. Nishirnura, R., Endo, Y., Tanabe, K., Ashitaka, Y. and Tojo, S. (1981) J. Endocrinol. Invest. 4, 349-355. Pierce, J.G. and Parsons, T.F. (1981) Annu. Rev. B&hem. 50, 465-495. Rademacher, T.W., Parekh, R.B. and Dwek, R.A. (1988) Annu. Rev. Biochem. 57, 785-838. Ryan, R.J., Keutman, H.T., Charlesworth, M.C., McCormick,
272 D.J., Milius, R.P., Calve, F.O. and Vutyavanich, T. (1987) Recent Rag. Harm. Res. 43, 383-429. Salomon, Y. (1979) Adv. Cyclic Nucleotide Res. 10, 35-55. Takasaki, S., Mizuochi, T. and Kobata, A. (1982) Methods Enzymol. 83, 263-268.
Thotakura, N.R. and Bahl, O.P. (1982) B&hem. Biophys. Res. Commun. 108, 399-405. Yamaahita, K., Mizuochi, T. and Kobata, A. (1982) Methods Enzymol. 83, 105-126.